ECG-ArrestNetIHCA Prediction Platform
Department of Cardiology · Central South University
Clinical Decision-Support Platform

ECG-ArrestNet: In-Hospital Cardiac Arrest Prediction from 12-Lead ECG

Department of Cardiology, Second Xiangya Hospital, Central South University, Changsha, China
This platform provides a deep learning-based tool for predicting in-hospital cardiac arrest (IHCA) from raw 12-lead ECG signals. Users may upload a recording or select a representative case, run the inference pipeline, and obtain a calibrated IHCA probability with feature-level attribution. The model integrates a multi-scale 1D-CNN backbone, lead-aware cross-attention, a BiLSTM temporal encoder, and gated fusion of 52 traditional ECG features with deep-learning representations.
Uploaded ECG data are processed under strict privacy protection and never stored or shared.
01 — Prediction Workspace

Analyze a 12-lead ECG

Provide an ECG via a built-in sample case, file upload, or manual feature entry. ECG-ArrestNet extracts electrophysiologic features, runs the full inference pipeline, and returns an IHCA probability with explainable contributions.

AECG Input
Sample case selected

Enter measured ECG features to run the prediction directly on the provided values.

12-LEAD ECG · 500 Hz · 10 sNormal Sinus Rhythm
Select a case and run the pipeline.
BInference & Results
Idle
✓
Stage 1
Multi-scale 1D-CNN
k=3/7/15/31 per lead
✓
Stage 2
Cross-Attention
8-head lead-aware
✓
Stage 3
BiLSTM Encoder
128 units, temporal
✓
Stage 4
Gated Fusion
52 traditional + DL
Run a prediction to view the IHCA risk score, feature contributions, and lead attention.
CSession Log
  • No analyses yet.
02 — About

Clinical decision support

ECG-ArrestNet is a clinical decision-support tool designed to assist clinicians in identifying patients at elevated risk of in-hospital cardiac arrest.

HHospital

Second Xiangya Hospital of Central South University — a tertiary referral center providing the clinical context and ECG data for ECG-ArrestNet development and validation.

AArchitecture

Multi-scale 1D-CNN backbone, 8-head lead-aware cross-attention, BiLSTM temporal encoder, and gated fusion of 52 traditional ECG features with deep-learning representations.

PPrivacy

Uploaded ECG data are processed under strict data-protection safeguards and are never stored, shared, or used for any other purpose.

How to use this platform

Select a representative case or upload a 12-lead ECG file (XML or CSV, 500 Hz, 10 s). Adjust clinical context (age, sex, care unit, ECG-to-event window) if available, then run the prediction. The output includes a calibrated IHCA probability, risk category, feature-level contributions, and a lead-attention heatmap.

03 — Clinical Notice

Important notice

Clinical decision-support — not a standalone diagnostic device

This platform is a clinical decision-support tool. It provides an IHCA risk probability to assist — not replace — clinical judgment. All predictions must be interpreted by qualified clinicians in conjunction with the patient's full clinical picture. In clinical use, predictions should be cross-validated against the hospital's electronic health record system. Uploaded signals are protected by strict data-security safeguards and are never transmitted or stored.

!Adjunctive use

ECG-ArrestNet is an adjunct to — not a substitute for — standard clinical assessment, vital-sign monitoring, and rapid-response protocols.

!Prevalence calibration

Risk probabilities should be interpreted in the context of local IHCA prevalence. PPV and NPV may differ from retrospective validation; recalibrate using Bayes' theorem for your setting.